Sie sind auf Seite 1von 21


discussions, stats, and author profiles for this publication at:

The Attachment Paradox: How Can So Many of

Us (the Insecure Ones) Have No Adaptive

Article in Perspectives on Psychological Science March 2010

DOI: 10.1177/1745691610362349


87 606

4 authors:

Tsachi Ein-Dor Mario Mikulincer

Interdisciplinary Center Herzliya Interdisciplinary Center (IDC) Herzliya


Guy Doron Phillip R Shaver

Interdisciplinary Center Herzliya University of California, Davis


Some of the authors of this publication are also working on these related projects:

Elaboration on Posttraumatic Stress Disorder Diagnostic Criteria: A Factor Analytic Study of PTSD
Exposure to War or Terror View project

Adult attachment theory and research View project

All content following this page was uploaded by Tsachi Ein-Dor on 14 September 2017.

The user has requested enhancement of the downloaded file.

Perspectives on Psychological

The Attachment Paradox : How Can So Many of Us (the Insecure Ones) Have No Adaptive Advantages?
Tsachi Ein-Dor, Mario Mikulincer, Guy Doron and Phillip R. Shaver
Perspectives on Psychological Science 2010 5: 123
DOI: 10.1177/1745691610362349

The online version of this article can be found at:

Published by:

On behalf of:

Association For Psychological Science

Additional services and information for Perspectives on Psychological Science can be found at:

Email Alerts:




Downloaded from at Interdisciplinary Center for on May 4, 2010

Perspectives on Psychological Science
5(2) 123-141
The Attachment Paradox: How Can So The Author(s) 2010
Reprints and permission:
Many of Us (the Insecure Ones) Have No
DOI: 10.1177/1745691610362349

Adaptive Advantages?

Tsachi Ein-Dor1,2, Mario Mikulincer1, Guy Doron1, and

Phillip R. Shaver3
The New School of Psychology, Interdisciplinary Center, Herzliya, Israel, 2Department of Psychology, Bar-Ilan University,
Ramat Gan, Israel, and 3Department of Psychology, University of California, Davis, CA

Bowlbys (1969/1982) attachment theory has generated an enormous body of research and conceptual elaborations. Although
attachment theory and research propose that attachment security provides a person with many adaptive advantages during all
phases of the life cycle, numerous studies indicate that almost half of the human species can be classified as insecurely
attached or insecure with respect to attachment. It seems odd that evolution left humans in this vulnerable position unless
there are some advantages, under at least some conditions, to anxious and avoidant attachment styles. We argue that a social
group containing members with different attachment patterns may be more conducive to survival than a homogeneous group
of securely attached individuals. In making this argument, we extend the scope of attachment theory and research by
considering a broader range of adaptive functions of insecure attachment strategies. We also present preliminary data to
support our argument.

attachment theory, social defense theory, insecure

Bowlbys (1969/1982, 1973, 1980) attachment theory postu- life span. For example, insecure adults (whether anxious or
lates an inborn behavioral system that emerged as an adaptation avoidant) report higher rates of relationship breakup (e.g.,
over the course of mammalian evolution. Because human Hazan & Shaver, 1987; Kirkpatrick & Hazan, 1994; Shaver
infants are born immature and require a long period of care and & Brennan, 1992) and more psychological problems (e.g.,
protection, they are equipped with a repertoire of behaviors that Mickelson, Kessler, & Shaver, 1997; Mikulincer, Florian, &
increase the likelihood that they will remain proximal to sup- Weller, 1993) than do secure adults. Nevertheless, research
portive others. That is, the attachment behavioral system indicates that almost half of living human beings in every age
presumably evolved because it increased the likelihood of group are insecure with respect to attachment, and the propor-
survival and eventual reproduction on the part of members of tion is higher in more disadvantaged (e.g., poorer, less socially
a species born with inadequate capacities for defense, locomo- stable) populations (Cassidy & Shaver, 1999, 2008; Mikulincer
tion, and feeding. & Shaver, 2007).
An enormous amount of research has been conducted since This set of facts raises questions concerning how such a high
Ainsworth and her students (e.g., Ainsworth, Blehar, Waters, & degree of insecurity could have survived evolutionary
Wall, 1978) first identified individual differences among winnowing over millions of years (Belsky, 1997, 1999; Belsky,
infants in the use of various attachment strategies (see Cassidy
& Shaver, 1999, 2008, and Mikulincer & Shaver, 2007, for
reviews). The major insecure attachment patterns, often labeled
Corresponding Author:
anxious and avoidant (e.g., Mikulincer & Shaver, 2007), are Tsachi Ein-Dor, The New School of Psychology, Interdisciplinary Center, P.O.
associated with relatively poor adjustment and, in some cases, Box 167, Herzliya, 46150, Israel
are associated with psychopathology at various phases of the E-mail:

Downloaded from at Interdisciplinary Center for on May 4, 2010 123

124 Ein-Dor, Mikulincer, Doron, and Shaver

Steinberg, & Draper, 1991; Chisholm, 1996; Simpson & 1969/1982). In such cases, the attachment system is activated
Belsky, 2008). In this article, we argue that secure and insecure and the individual is driven to reestablish actual or symbolic
attachment styles may both have unique adaptive advantages proximity to an external or internalized attachment figure until
(which increase inclusive fitness; see Hamilton, 1964) and felt security is attained. Bowlby (1969/1982, 1988) assumed
disadvantages (which decrease inclusive fitness). As a result, that although age and development result in an increased ability
when viewed from either an inclusive-fitness perspective or a to gain comfort from symbolic representations of attachment
group-selection perspective, groups that include individuals figures, no one of any age is completely free from reliance
with different attachment strategies may have advantages over on actual others. The attachment system therefore remains
homogenous groups of relatively secure individuals. active over the entire life span, as indicated by adults tendency
Specifically, we argue that relatively secure individuals tend to seek proximity and support when threatened or distressed
to remain emotionally stable in the face of threats and can (Hazan & Zeifman, 1999).
calmly and efficiently coordinate group members problem- Many studies have shown that the attachment system is
solving efforts. But this behavior may be counterproductive indeed active during adulthood and that it affects many aspects
in certain circumstances. For example, when a serious danger of psychological and social functioning (see relevant chapters
arises, the optimism inherent in the secure-base script, as well in Cassidy & Shaver, 2008; Mikulincer & Shaver, 2007). For
as the secure persons concern for staying proximal to relation- instance, Mikulincer and colleagues (e.g., Mikulincer,
ship partners, may reduce a secure individuals ability to recog- Birnbaum, Woddis, & Nachmias, 2000; Mikulincer, Gillath,
nize the gravity of the threat or respond to it effectively in time & Shaver, 2002) showed that perceived threats generally cause
to facilitate a safe escape. In such cases, anxious individuals adults to activate mental representations of attachment figures
hypervigilance to threats may benefit members of a group by as a means of coping and regulating emotions. Study partici-
reacting to early or ambiguous cues of danger and alerting pants reacted to threats with heightened mental accessibility
everyone to the threat. Avoidant individuals concern with of the names of their supportive attachment figures in both
self-preservation may facilitate their discovery of an escape lexical decision tasks and Stroop color-naming tasks. As com-
route that others can use, even if it was not intended for the oth- pared with neutral subliminal priming, subliminal priming with
ers benefit. In other words, what are usually interpreted as threat words (e.g., failure, death, or illness) led to both faster
maladaptive attachment strategies may, under some conditions, identification of attachment figures names in a lexical decision
be beneficial for survival of an individual and members of task and slower color-naming times for attachment figures
the individuals group (all of whom, in the environment of names in a Stroop task. Fast lexical decision times and slow
evolutionary adaptedness, would often have been genetically color-naming responses were interpreted as indications of
related). heightened activation of mental representations of attachment
In contrast to Belsky and colleagues, who have attempted to figures in threatening contexts.
link insecure attachment patterns to particular mating strate- Bowlby (1969/1982) listed some of the major situations or
gies, we suggest that the adaptive advantages of insecure stimuli that activate the attachment system in early childhood:
attachment patterns may be that they promote the survival of
individuals in a group rather than directly promoting reproduc- Observation of any one child during his second and third years
tion. Of course, there would be no reproduction without sur- of life, when attachment behavior is most evident, shows that
vival. Our analysis is not meant to substitute for those of such behavior varies enormously in activation, form, and
Belsky and his colleagues; it is meant to supplement them. But intensity. At one moment the child is content to explore his
we do provide reasons for viewing survival benefits of insecure surroundings; at the next he is searching for his mother despe-
attachment as more important and more likely than direct rately.... Every mother knows that a child who is tired, hungry,
reproductive benefits. cold, in ill health or in pain is likely to be especially mummy-
ish. (p. 311)

Attachment-System Functioning Under normal circumstances, this activation is manifested in

According to Bowlby (1969/1982, 1973, 1980), human beings the actual seeking of proximity to attachment figures. How-
and many nonhuman primates are born with an innate psycho- ever, there are cases in which these behaviors may be inhibited
biological system (the attachment behavioral system) that by the absence of attachment figures or by other personal and
motivates them to seek proximity to significant others (attach- contextual factors. In such cases, thoughts about proximity to
ment figures) in times of need as a way of protecting them- attachment figures may still be mentally active, and representa-
selves from threats and alleviating distress. The goal of the tions of these figures may still influence behavior (Mikulincer
system is objective protection or support and the concomitant & Shaver, 2003).
subjective sense of safety or security (which Sroufe & Waters, In addition to describing the normative activation and oper-
1977, called felt security). This double-sided (objective and ation of the attachment system, Bowlby (1973, 1988) and his
subjective) goal is made salient when a person encounters research collaborator Ainsworth (e.g., Ainsworth et al., 1978)
actual or symbolic threats and notices that an attachment figure focused on salient individual differences. Although the attach-
is not sufficiently near, interested, or responsive (Bowlby, ment system is active in every individual, there are individual

124 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 125

differences in the extent to which attachment behaviors are attachment) have developed alternative ways of coping with
effective in mitigating distress and attaining felt security. When threats and stress. Avoidant attachment is organized around
people can attain support and protection when needed, they feel deactivating strategies of affect regulation, which involve
confident about the value of proximity seeking as a coping deemphasizing threats and trying to cope with them alone,
strategy, are able to trust other peoples availability and will- without seeking help or support from other people (e.g., Fraley
ingness to provide help and support, and are increasingly able & Shaver, 1997; Kobak, Cole, Ferenz-Gillies, & Fleming,
to build resources and skills that allow them to deal autono- 1993; Mikulincer & Orbach, 1995; Shaver & Mikulincer,
mously with stress. However, when people fail to attain support 2002). Anxious attachment is organized around hyperactivat-
and protection from attachment figures, worries and insecuri- ing strategies of affect regulation, which involve overempha-
ties about proximity seeking and others goodwill are formed, sizing threats and becoming very emotional and intrusive or
distress is intensified, and defensive strategies other than con- insistent in attempts to gain protection and support from other
fident proximity seeking are developed. Main (1990) labeled people (Cassidy & Kobak, 1988; Feeney & Noller, 1990;
these alternatives secondary attachment strategies, and later Mikulincer et al., 2000; Mikulincer, Orbach, & Iavnieli,
theorists believe they involve either hyperactivation or deacti- 1998; Shaver & Mikulincer, 2002). Both of the major insecure
vation of the attachment system (Cassidy & Berlin, 1994; attachment strategies tend to cause difficulties in relationships
Cassidy & Kobak, 1988; Mikulincer & Shaver, 2003, 2007). and are often the foci of individual and couple psychotherapy
This analysis of individual differences in attachment strate- (Cassidy & Shaver, 1999, 2008).
gies has influenced personality and social psychologists who How do these individual differences in attachment style
study attachment in adolescent and adult samples to focus on arise? According to Bowlby (1973) and Ainsworth et al.
individual differences in attachment style (see review by (1978), they are shaped by interactions with ones attachment
Mikulincer & Shaver, 2007). Adolescent and adult attachment figures across the life span, but especially in childhood. Inter-
styles are generally conceptualized as regions in a two- actions with attachment figures who are available and respon-
dimensional space (Brennan, Clark, & Shaver, 1998; Fraley sive contribute to a core dispositional sense of attachment
& Waller, 1998). One dimension, attachment-related avoid- security. However, when attachment figures are not reliably
ance, reflects the extent to which a person distrusts relationship available and supportive, a sense of security is not attained and
partners goodwill, strives to maintain independence, and relies secondary strategies of affect regulation come into play.
on deactivating strategies for dealing with dangers and threats. According to Bowlby (1973), interactions with rejecting and
The second dimension, attachment-related anxiety, reflects the unsupportive attachment figures push a child toward one or
extent to which a person worries that a relationship partner will another attachment strategy, which is characterized by particu-
not be available in times of need and hyperactivates the attach- lar mental representations of self and attachment figures called
ment system in an attempt to gain the partners attention, care, internal working models (Bretherton & Munholland, 2008).
and love (Mikulincer & Shaver, 2003). Attachment security is These working models allow a person to predict future interac-
indicated by low scores on measures of both insecurity dimen- tions with attachment figures and activate generally reliable
sions. Scores in the other regions of the space indicate attach- strategies for interacting with them that do not require rethink-
ment insecurity of various kinds, each with its own emotional ing or initiating a particular set of strategic actions each time
and behavioral correlates. Throughout this article, we refer to they seem relevant to a situation.
people with secure, anxious, and avoidant attachment styles, Research has shown that these individual differences are
or people who are relatively anxious or avoidant. Although the transmitted from parents to children in what van IJzendoorn
convenient categorical shorthand (secure, anxious, and avoi- (1995) called the intergenerational transmission of attach-
dant) can mistakenly foster typological thinking, we will ment (p. 387). High rates of concordance (between 60% and
always be referring to fuzzy regions in a two-dimensional 85%) have been found, for example, between a mothers state
spacea space in which research participants are continuously of mind with respect to attachment (as measured by the Adult
rather than categorically distributed. Attachment Interview; Hesse, 2008) and her childs degree of
Research supports the claim that a persons attachment style security or insecurity in her presence in a laboratory assessment
is related to his or her ways of coping with threats. For exam- procedure called the Strange Situation (Ainsworth et al., 1978;
ple, people who score low on both attachment anxiety and Benoit & Parker, 1994; Fonagy, Steele, & Steele, 1991; Hesse,
avoidance (i.e., those who are relatively secure with respect 2008). According to Main, Kaplan, and Cassidy (1985), the
to attachment) are commonly observed to cope with threats quality of parentchild interactions mediates this intergenera-
by effectively seeking support from others (Mikulincer & tional transmission of attachment. That is, parents attachment
Florian, 1995, 1998; Mikulincer et al., 1993), possess a strong working models shape their caregiving behavior and affect
sense of self-efficacy (Collins & Read, 1990; Mikulincer & their ability and willingness to provide a safe haven and secure
Florian, 1995), generally trust other people (Bartholomew & base for their child, which in turn contributes to the childs
Horowitz, 1991; Collins & Read, 1990), and perceive the world attachment security. However, van IJzendoorns (1995) meta-
as a relatively safe place (Mikulincer & Shaver, 2007). analysis of 10 studies revealed what he called a transmission
People who score high on attachment anxiety and/or avoid- gap. Although there was a sizeable mean effect size linking a
ance (i.e., those who are relatively insecure with respect to secure parental state of mind with respect to attachment and

Downloaded from at Interdisciplinary Center for on May 4, 2010 125

126 Ein-Dor, Mikulincer, Doron, and Shaver

sensitive responsiveness to childrens needs and signals during Although more than 2 billion people worldwide apparently
parentchild interactions, much of the association between par- have an insecure attachment style, theoretical proposals regard-
ents and childs attachment status seemed to occur through ing possible adaptive functions of such patterns are scarce (e.g.,
processes other than the quality of parentalchild interactions, Belsky, 1997, 1999; Belsky et al., 1991; Simpson & Belsky,
at least as this has been measured to date. 2008). By far, most of the research has focused on the disad-
The transmission gap opened the door to possible genetic vantages of secondary attachment strategies. For example,
explanations of intergenerational transmission of attachment research examining individual differences in attachment orien-
patterns, which challenged Bowlbys (1969/1982) almost exclu- tations has linked avoidant attachment with fear of intimacy
sive emphasis on the importance of early social experiences. and a tendency to maintain distance in close relationships, pes-
And of course, genetic transmission could have played a role simistic beliefs about relationships, proneness to sexual infide-
even if there had been no transmission gap, because the similar- lity in relationships (Brennan, Shaver, & Tobey, 1991), and a
ity between a parents and his or her childs attachment security high rate of relationship dissolution (see Mikulincer & Shaver,
could have been based partly on shared genes. Behavioral 2007, for a review). Avoidant individuals disapprove of and
genetic studies that have assessed concordance of attachment avoid self-disclosure (e.g., Mikulincer & Nachshon, 1991), are
patterns in monozygotic and dizygotic twins have indicated that susceptible to sudden and erratic religious conversions (e.g.,
genetic factors can explain between 14% and 40% of the var- Kirkpatrick & Shaver, 1990), and are judged by their peers to
iance of attachment patterns at various phases of the life span be unusually hostile (e.g., Kobak & Sceery, 1988). They prefer
(e.g., Brussoni, Jang, Livesley, & MacBeth, 2000; Crawford to work alone and use work as an excuse for avoiding close
et al., 2007; Finkel & Matheny, 2000; OConnor & Croft, relationships (e.g., Hazan & Shaver, 1990). They are more
2001). Studies exploring possible molecular genetic markers likely than less avoidant individuals to report that they have
associated with attachment patterns have shown that different or had poor relations with their parents while attending college
attachment patterns are somewhat related to particular genetic (e.g., Rothbard & Shaver, 1994; Shaver & Clark, 1994). They
alleles (e.g., Donnellan, Burt, Levendosky, & Klump, 2008; attempt to cope with stress by ignoring or denying it (e.g.,
Gillath, Shaver, Baek, & Chun, 2008). However, other studies Dozier & Kobak, 1992) or by using alcohol and other sub-
have failed to find a strong, direct genetic contribution to attach- stances to reduce tension (e.g., Brennan et al., 1991). Following
ment patterns (e.g., Bakermans-Kranenburg & van IJzendoorn, stressful periods, they are more likely than nonavoidant indi-
2004) and have instead found evidence for an interaction viduals to exhibit psychosomatic symptoms (e.g., Mikulincer
between genetic and parenting influences. et al., 1993). They are somewhat pessimistic and cynical about
long-term relationships (e.g., Carnelley & Janoff-Bulman,
1992) and are less compassionate and altruistic than their less
The Attachment Theory Paradox avoidant peers (Mikulincer & Shaver, 2005).
One of the most frequently replicated findings in attachment Individuals who are anxious with respect to attachment also
research concerns the distribution of attachment orientations contend with many problems and disadvantages, such as an
in infancy, childhood, adolescence, and adulthood. Approxi- obsessive preoccupation with romantic partners interest, loy-
mately 33% of infants, children, adolescents, and adults exhibit alty, and responsiveness; jealousy; a tendency toward fear, anxi-
insecure attachment patterns. For instance, using the Strange ety, and loneliness (even when involved in a relationship); low
Situation testing procedure for infants, numerous researchers and unstable self-esteem (e.g., Collins & Read, 1990; Feeney
in the United States (see reviews in Grossmann, Grossmann, & Noller, 1990); and a higher than usual rate of relationship dis-
& Waters, 2005), Germany (e.g., Beller & Pohl, 1986; solution (e.g., Hazan & Shaver, 1987). Such individuals are
Grossmann, Grossmann, Huber, & Wartner, 1981), Japan excited about leaving home and going to college, but they
(e.g., Miyake, Chen, & Campos, 1985), Sweden (e.g., Lamb, become socially dissatisfied and lonely after the first semester
Hwang, Frodi, & Frodi, 1982), the Netherlands (e.g., van IJzen- (e.g., Hazan & Hutt, 1993). In addition, they prefer to work with
doorn, Goossens, Kroonenberg, & Tavecchio, 1985), Indonesia others but feel unappreciated and misunderstood at work and
(e.g., Zevalkink, Riksen-Walraven, & Van-Lieshout, 1999), tend to daydream about success and slack off after receiving
and Israel (e.g., Sagi et al., 1985) have found that, on average, praise (e.g., Hazan & Shaver, 1990). They become very emo-
the distribution of infants is roughly 65% secure, 20% avoidant, tional under stress and are forced to use ineffective emotion-
and 15% anxious (see review in van IJzendoorn & Sagi, 1999). focused coping strategies (e.g., Mikulincer et al., 1993). They
In a meta-analysis of almost 2,000 Strange Situation classifica- report more physical and psychological symptoms (e.g., Fiala,
tions from 32 samples in eight countries, van IJzendoorn and 1991; Hazan & Shaver, 1990). They worry about rejection dur-
Kroonenberg (1988) showed that cross-cultural differences in ing interactions (Tidwell, Reis, & Shaver, 1996) and indiscrimi-
attachment distributions were relatively small. Studies con- nately self-disclose too much, too soon (e.g., Mikulincer &
ducted with adolescents and adults have found similar distribu- Nachshon, 1991). They tend to be argumentative, intrusive, and
tions of attachment patterns using both self-reports scales (e.g., overly controlling (e.g., Kunce & Shaver, 1994).
Feeney & Noller, 1990; Hazan & Shaver, 1987; Mikulincer, Both kinds of insecure attachment have been associated
Florian, & Tolmacz, 1990) and the Adult Attachment Interview with various forms of psychopathology, such as depression
(see Hesse, 2008, for a review). (e.g., Carnelley, Pietromonaco, & Jaffe, 1994), anxiety (e.g.,

126 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 127

Bartholomew, 1990; Riskind et al., 2004; Safford, Alloy, In other words, Belsky and associates proposed that the
Crossfield, Morocco, & Wang, 2004; Williams & Riskind, potential to develop different attachment patterns evolved
2004), obsessive-compulsive disorder (e.g., Doron & Kyrios, because they promote reproductive fitness under certain
2005), and eating disorders (e.g., Friedberg & Lyddon, ecological conditions. Whereas classical attachment theory
1996). Conversely, a secure attachment style seems to buffer emphasized the survival function of the attachment system,
a person against various risk factors (Mikulincer & Shaver, Belsky (1999) argued that modern evolutionary thinking
2007; Thompson, 2008). Self-reports of attachment security are focuses on reproductive fitness and that insecure attachment
associated with increased perceptions of self-efficacy, positive strategies may be reproductively advantageous under certain
affect, and reliance on problem-solving coping strategies for conditions.
dealing with personal and interpersonal stressors (e.g., This argument is based on Symonss (1987) suggestion that
Birnbaum, Orr, Mikulincer, & Florian, 1997; Collins & Read, the adaptive functions of all morphological, physiological, and
1990; Lussier, Sabourin, & Turgeon, 1997; Mikulincer, 1998; psychological attributes are ultimately in the service of genetic
Mikulincer & Florian, 1998). reproduction. However, Symons did not argue that all of these
Although a great deal of research has documented the disad- attributes evolved because they directly promoted reproduc-
vantages of high levels of attachment anxiety or avoidance, a tion. Although the final goal is reproduction, many morpholo-
closer examination of these insecure tendencies suggests possi- gical, physiological, and psychological attributes evolved
ble adaptive advantages under special conditions. Indeed, evo- because they solved a specific problem of survival. They
lutionary theory suggests that anything as pervasive in the promote reproduction by enabling an organism to reach a life
population as insecure attachment is likely to have an adaptive stage in which successful reproduction and rearing of offspring
function. In his analysis of variability in scores on the Big Five are possible.
personality traits, for example, Nettle (2006) argued that such Belsky argued that humans evolved in a way that allows
variability can be understood in terms of trade-offs among dif- them to modify their reproductive behavior (i.e., mating
ferent fitness benefits and costs. According to Nettle (2006), strategies) in the service of fitness and in accord with ecologi-
behavioral alternatives can be considered as trade-offs, with cal conditions. For instance, when resources are limited and
a particular trait producing not unalloyed advantage but a mix- unpredictable, parents may unintentionally provide insensitive
ture of costs and benefits such that the optimal value for fitness and unreliable care to their children. In such cases, children
may depend on very specific local circumstances (p. 625). may develop avoidant attachment strategies. When they
Thus, although low levels of extraversion and agreeableness become adolescents or adults, they may pursue a reproductive
and high levels of neuroticism are typically viewed as imposing strategy that emphasizes mating (and procreation) over parent-
important adjustment costs, they can also make important con- ing. That is, by encouraging quantity over quality of offspring,
tributions to fitness, which preserves genetic variations in these avoidant attachment may indirectly promote reproductive fit-
traits. In terms of Nettles analysis, introversion has fitness ben- ness in an unstable, insecurity-arousing environment. Belsky
efits related to avoiding physical risks and maintaining family (1999) made the following comment:
stability, neuroticism may be beneficial by causing a person to
be vigilant with respect to danger, and low agreeableness may Since the risks to progenys eventual reproductive success were
be beneficial in making a person less subject to others dishon- great and uncontrollable [when the flow of resources was
esty while giving the person a selfish advantage. Following this chronically low or unpredictable], the way to ensure that some
line of argument, it seems worthwhile to consider possible [children] would survive to reproduce would involve the pro-
fitness benefits of attachment anxiety and avoidance. duction of many offspring, even if they were not going to be
well cared for. (p. 154)

Belskys Evolutionary Analysis of Attachment The adaptive advantage of avoidant people is that they pursue
Insecurity disproportionately self-serving, opportunistic sexual matings
and therefore are likely to have multiple mates and many chil-
Belsky and colleagues (Belsky, 1997, 1999; Belsky et al., 1991;
drensome of whom, even if poorly cared for, may go on to
see also Chisholm, 1996; Simpson & Belsky, 2008) were the
reproduce in turn.
first to address the possibility that there are potential adaptive
With regard to attachment anxiety, Belsky argued that
benefits of insecure attachment styles under particular environ-
because whining dependency (1999, p. 156) is costly to par-
mental conditions. For example, they stated as follows:
ents, this pattern could have evolved only if it offered a biolo-
Attachment behavior would not have evolved if it had only gical pay-off for the parents as well as the child. He reasoned as
functioned to protect the child and thereby to promote survival, follows:
because survival per se is clearly not the goal of natural
selection. . . . Thus variations in attachment security that The capacity for developing resistant [i.e., anxious] attach-
Bowlbys theory so clearly anticipated evolved to serve repro- ments evolved as a means of fostering indirectly reproductive
ductive fitness goals in an ecologically sensitive manner. helper-at-the-nest behavior. That is, by inducing helpless
(Belsky, 1999, p. 142) dependency in a child, inconsistently responsive parenting

Downloaded from at Interdisciplinary Center for on May 4, 2010 127

128 Ein-Dor, Mikulincer, Doron, and Shaver

evolved to promote a reproductive strategy designed to range. Once a snake is perceived to be dangerous and within
facilitate the direct reproductive success of kin (especially striking distance, this information causes an increase in auto-
parents), and thereby the indirect reproductive success of the nomic arousal, which is part of a flight response that would
resistant [anxious] individual. (p. 156) have protected early humans from deadly snake bites.
The attachment system is also triggered by a fairly narrow
The adaptive advantage of anxious individuals, according to range of inputs (which Bowlby, 1973, p. 138, called natural
this argument, is that rather than leave the family home to clues of danger), which produce a limited set of behavioral out-
reproduce, they may stay and help other family members give puts (proximity maintenance, signaling a desire to be picked up,
birth and raise children, thereby increasing the likely survival and hugging and clinging). These cues continue to activate the
and successful reproduction of kin with whom they share attachment system across the life span (e.g., Mikulincer et al.,
genes. 2000; see Mikulincer & Shaver, 2007, for a review). Thus, sur-
Recently, Del-Giudice (2009) broadened Belskys theory vival rather than immediate reproduction seems to be the major
and suggested an integrated evolutionary model of the develop- reason for the evolution of an attachment behavioral system.
ment of attachment and human reproductive strategies. He Also relevant to evaluating Belskys and related theories are
argued that early psychosocial stress and attachment insecuri- studies of sex differences. According to evolutionary psychol-
ties cause a developing individual to pursue reproductive ogy, differences between males and females are expected in
strategies focused on mating effort and early reproduction domains in which they have recurrently faced different adap-
(rather than investing in long-term relationships and parental tive challenges. Conversely, in domains in which the sexes
investment, which is the strategy associated with early attach- have faced the same adaptive problems, no sex differences are
ment security). Due to sex differences in trade-offs between mat- expected (Buss, 1995). In the course of evolution, men and
ing and parenting, insecure males tend to adopt avoidant women have faced different problems pertaining to mating
strategies and insecure females tend to adopt anxious strategies. (Buss, 1999). Therefore, we would expect there to be sex dif-
Studies that support Del-Giudices (2009) analysis show that ferences in mating strategies and in attachment strategies if
avoidant men tend to exhibit what Bartholomew and Horowitz attachment behaviors evolved because they directly promoted
(1991) called dismissing avoidance, whereas avoidant women sexual reproduction. In fact, however, although there are sex
tend to exhibit fearful avoidance. But there are many other stud- differences in mate preferences (Buss, 1989), courting strate-
ies that found no gender differences in scores on measures of gies (Buss, 1988a, 1988b; Tooke & Camire, 1991), jealousy
adult attachment (see Mikulincer & Shaver, 2007, for a review). (Buss, Larsen, Westen, & Semmelroth, 1992), and mate-
Although Belsky took an important first step in attempting guarding tactics (Buss, 1988a; Flinn, 1988), there are not reli-
to provide an adaptationist analysis of insecure attachment, able, general sex differences in attachment strategies (although,
there are reasons to challenge his basic assumptions. First, as mentioned already, there may be gender differences in the
recent theoretical analyses and empirical studies in evolution- prevalence of dismissing and fearful forms of avoidance).
ary psychology suggest that psychological mechanisms exist With regard to jealousy, for example, evolutionary psychol-
in their current form because they helped to resolve specific ogists predicted that the sexes would differ in the kinds of
problems of individual survival or reproduction (Buss, 1995; events that activate this intense emotion (Daly, Wilson, &
Tooby & Cosmides, 1990). The fear system, for example, Weghorst, 1982; Symons, 1979). Because of concealed ovula-
including the fight or flight response, is an example of a psy- tion and the fact that fertilization and gestation occur internally
chobiological system that evolved because it helped resolve a in women, men have faced an adaptive problem not faced by
specific problem threatening individual survival. Fighting or womenhaving less than 100% certainty of biological parent-
fleeing from a predator probably did not, and still does not, hood. The reproductive threat for men comes from the possibil-
serve immediate reproductive goals, but it may make ity of sexual infidelity on the part of a female mate. In species
reproduction possible by increasing the chances of survival to such as ours, in which a man sometimes invests heavily in a
reproductive age. woman and her children, the womans certainty of genetic par-
According to Bowlbys (1969/1982) theory, the attachment enthood is not compromised if her mate has sex with other
system is closely related to the fear system: It is activated by women. However, the woman may risk the loss of her mates
threats, and the behavioral output of the system is generally time, attention, commitment, involvement, protection, and
adaptive because it increases the likelihood of survival. If resourcesresources that can be diverted from her and her
human infants were not interested in and able to signal urgent children to another woman. For these reasons, evolutionary
needs and desires for help and protection, they would die long psychologists predicted and have generally found that the
before sexual reproduction became a primary interest. Con- triggers of jealousy in men would be related to the sex act per
sider, for purposes of comparison, the disposition to fear snakes se, whereas for women the main issue would be loss of a mans
(Marks, 1987). It presumably exists in humans and nonhuman commitment and investment (but see Harris, 2003, for a critical
primates because it contributed to the solution of a specific evaluation of this literature).
problem of survival in primates ancestral environments. The Lack of consistent sex differences in attachment patterns
fear of snakes is triggered only by a narrow range of inputs, does not fit with the thrust of Belsky and colleagues argu-
such as long, sliding organisms perceived to be within striking ments, because differences between males and females would

128 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 129

be expected if attachment styles (or the potential for insecure reproductive efforts of kin? Hamilton proposed that the indi-
attachment styles under certain ecological conditions) evolved viduals total (inclusive) fitness depends on his or her own
directly, or mainly, to promote reproduction. Instead, the gen- reproductive output plus the total reproductive output of kin
eral lack of consistent sex differences in attachment style are who share the individuals genes. Moreover, in 1975, Hamilton
consistent with Bowlbys (1969/1982) idea that all human realized that many social traits (e.g., altruism) may be selec-
beings, male and female alike, have faced the adaptive problem tively disadvantageous at the individual level but might have
of being born immature and needing care and protection from evolved nevertheless because families and groups containing
attachment figures. members with these social traits could contribute more to the
It is important to note, however, that some studies have total gene pool than other kinds of families and groups
found an association between avoidant attachment and short- (Hamilton, 1975). In other words, between-group selection
term mating strategies (Schmitt, 2005). Likewise, some studies may favor certain social traits, even though they are disadvan-
have found that anxiously attached people continue to perceive tageous at the individual level (Sober & Wilson, 1998; Wilson,
their parents as their primary attachment figures later in life as Vugt, & OGorman, 2008; Wilson & Wilson, 2007; but see
compared with secure and avoidant people (Feeney, 2004). Dawkins, 1994, for an early critical evaluation of multilevel
This finding might seem to support Belskys (1999) suggestion selection theory). For our purposes, the evolution and
that anxious adults are more likely to be helpers in the nest. maintenance of multiple attachment styles could be considered
In his words, It is usually only where there is some biological a consequence of either inclusive fitness or multilevel selection
payoff for the parent as well as for the child that we should or a combination of the two.
expect a parentally influenced and costly pattern of child In the course of evolution, humans lived in small, highly
functioning to evolve (p. 156). However, other studies interactive groups of kin, and humans are unique among mam-
(B.D. Carpenter, 2001; Cicirelli, 1993; Crispi, Schiaffino, & mals in the duration and complexity of the social relationships
Berman, 1997; Eberly & Montemayor, 1998, 1999; Townsend they form. Because social solutions to adaptive challenges
& Franks, 1995) indicate that anxiously attached people are were so crucial for human survival, many of our psychological
less concerned with the well-being of their parents and show mechanisms undoubtedly evolved to support this aspect of
lower willingness to care for them. In other words, anxiously human existence (Buss, 1995). Brewer and Caporael (1990),
attached people may be in the nest longer, but not necessarily as well as other scholars (e.g., Alexander, 1987; Axelrod,
because they are inclined to be helpers in the nest. 1984; Cosmides, 1989; Gazzaniga, 2008), have argued that liv-
ing in cooperative groups was the primary survival strategy for
humans. This raises the possibility that different attachment
Looking for Other Adaptive Advantages of
orientations, or the facultative potential to develop such orien-
Secondary Attachment Strategies tations under certain environmental conditions (Belsky, 1999;
Hundreds of individual-difference studies have not found adap- Chisholm, 1996; West-Eberhard, 2003), might have contribu-
tive advantages of insecure attachment styles (Mikulincer & ted to inclusive fitness because they are advantageous for mul-
Shaver, 2007), except possibly for the ones that Belsky and his tiple members of a group, extended family, or small tribe. All
colleagues focused on. The reason for failing to see adaptive of the major attachment patternssecure, anxious, and avoi-
benefits of insecurity may be twofold. First, contemporary dantmay promote inclusive fitness in their own ways. As
researchers have focused on what counts as individual mental we explain more fully below, the secure pattern may be associ-
health in a modern society, more or less assuming that this state ated with a balanced, sensible, somewhat altruistic coordina-
of mental health has always been an adaptive advantage. There tion and leadership of group responses to challenges and
has not been much reason to consider adaptive advantages of threats. The avoidant pattern may be associated with quick,
what is currently viewed as maladaptive behavior in modern, independent responses to threat, which may at times increase
industrialized, and urban societies. Second, studies motivated the survival chances of group members by solving the survival
by attachment theory may have concentrated too exclusively problem or demonstrating ways to escape it. The anxious pat-
on one level of analysis, that of individual well-being, mating, tern may be associated with sensitivity and quick detection of
and reproduction. Here, we consider the possibility that some dangers and threats, which alert other group members to danger
of the evolutionary advantages reside at the group level and and the need for protection or escape.
concern survival rather than mating and reproduction (which Our analysis is based on the following three proposals:
of course depend on survival). As Bowlby (1969/1982) said,
From a new viewpoint a familiar landscape can sometimes 1. Secure and insecure attachment styles may have both
look very different (p. 1). unique adaptive advantages (which increase inclusive fit-
We base our argument on Hamiltons (1964) kin selection ness) and disadvantages (which decrease inclusive fitness;
theory while also taking Sober and Wilsons (1998) multilevel see Hamilton, 1964, for other examples of such mixtures).
selection theory into account. Hamilton tried to overcome one 2. Different attachment patterns, by virtue of their associated
of the biggest stumbling blocks for Darwins theory of evolu- internal working models and action tendencies or
tion: In the evolutionary struggle for reproductive fitness, why behavioral scripts, may have different benefits for group
do some organisms forgo reproduction while assisting the members under threatening conditions. Specifically,

Downloaded from at Interdisciplinary Center for on May 4, 2010 129

130 Ein-Dor, Mikulincer, Doron, and Shaver

individuals who are relatively high in attachment anxiety and organized by anticipation and simulation (Cross & Markus,
may be characterized by what we will call sentinel reac- 1994).
tions and scripts, and individuals who are relatively high The same logic applies to attachment orientations, which
on avoidant attachment may be characterized by what we include self-schemas as parts of internal working models
will call rapid fightflight reactions and scripts. (Bretherton & Munholland, 2008; Waters, Rodrigues, &
3. An extended family or tribal group that is heterogeneous Ridgeway, 1998; Waters & Waters, 2006). Although the exist-
with respect to attachment orientations, like groups that are ing literature on attachment tends to emphasize adaptive
heterogeneous in other respects (see Nettle, 2006), may be aspects of secure working models or scripts and maladaptive
more effective in dealing with threats and survival prob- aspects of anxious or avoidant working models, we are espe-
lems than a homogenous group or society. cially interested here in cases in which a secure attachment
style has disadvantages for members of a group or community
These three propositions, when fleshed out in the next section, and cases in which insecure attachment styles contribute to
constitute what we call social defense theory (SDT). group members survival and promote inclusive fitness.

SDT Possible Disadvantages of Security Attachment

SDT postulates that each attachment orientationsecure, Research on adult attachment has shown that individuals with
anxious, or avoidanthas both unique adaptive advantages for relatively low scores on attachment anxiety and avoidance
increasing the inclusive fitness of individuals in a group and (i.e., those who are relatively secure with respect to attachment)
unique disadvantages that decrease inclusive fitness (see provide many advantages to the groups in which they partici-
Nettle, 2006, for related ideas about variations in personality). pate. For example, they are generally better than less secure
According to the theory, each attachment orientation includes a individuals at leading and coordinating group activities
working model, or script, that is especially available under (Davidovitz, Mikulincer, Shaver, Ijzak, & Popper, 2007; Rom
threatening conditions. This script renders an individual likely & Mikulincer, 2003). Also, they spend more time than their less
to act in a manner consistent with his or her attachment style. secure counterparts in groups, engage in more group activities,
According to Markus (1977), every individual perceives the and work more effectively with other group members when sol-
world in ways that are organized by self-schemas, which ving problems and resolving interpersonal conflicts (Rom &
include mental representations of the persons special abilities, Mikulincer, 2003; Smith, Murphy, & Coats, 1999). According
achievements, and preferences. Individuals with self-schemas to Mikulincer and Shavers (2003, 2007) literature reviews,
in a particular domain use them to make quick and confident these advantages stem from a sense of security rooted in past
judgments in that domain, adapt flexibly to different supportive experiences with attachment figures. This sense of
information-processing goals related to the domain, and accu- security is closely associated with core beliefs, such as the
rately retrieve information relevant to the domain (e.g., S.L. belief that the world is a safe place, especially when significant
Carpenter, 1988; Kanagawa, Cross, & Markus, 2001; Markus, others are present. These optimistic, comforting mental repre-
1977, 1980). People are chronically sensitive to schema- sentations promote self-soothing reappraisals of threats, which
relevant stimuli and pay close attention to information perti- help secure individuals perform better than insecure individu-
nent to the schemas domain (Markus & Sentis, 1982). As a als in many challenging situations.
result, they are more ready than aschematic individuals to apply However, feelings of security do not always reflect actual
relevant abilities and skills when the schema is activated (see physical security. In most everyday situations, a sense of secu-
Markus, 1983; Markus, Cross, & Wurf, 1990; Markus & Wurf, rity is beneficial because it reduces anxiety and allows a person
1987, for reviews). to focus fully on the task at hand. This is probably one reason
Moreover, people who possess a self-schema in a particular why relatively secure individuals perform better than insecure
domain are better able to predict their future behavior in that ones at work (Hazan & Shaver, 1990). In times of actual dan-
domain (Markus, Crane, Bernstein, & Siladi, 1982). Markus ger, however, the sense of security may be maladaptive if it
and Nurius (1986) called these future-oriented self- causes a person to postpone recognizing the emerging threat.
conceptions possible selves. Possible selves are the elements Focusing on an ongoing project irrespective of mounting dan-
of a self-system that allow an individual to simulate the neces- ger may interfere with taking effective action in an emergency.
sary steps and strategies for accomplishing a certain goal (see A quick fight-or-flight response is sometimes necessary to
Markus et al., 1990; Markus & Ruvolo, 1989, for reviews). avert disaster, and being good at self-soothing and generating
Possible selves enable a person to focus attention on specific, optimistic appraisals in such cases may be counterproductive
task-relevant thoughts and feelings and to organize action for both the individual and the group.
(Inglehart, Markus, & Brown, 1989), building a bridge between Indeed, Bowlby (1973) observed that attachment behavior
the current situation and a future outcome (Oyserman & frequently takes precedence over escape in many species.
Markus, 1990). The more vivid and elaborate the possible Later, Mawson (1978, 1980, 2005) showed that the typical
selves, the better the performance, because many of the rou- human response to danger is to seek the proximity of familiar
tines required for performance are already activated, engaged, people and places, even if this means remaining in or even

130 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 131

approaching a dangerous situation (see also Baker & Chapman, Studies indicate that most people tend to remain in a disaster
1962; Henderson, 1977; Kinston & Rosser, 1974). Secure indi- area (Cantril et al., 1940; Quarantelli & Dynes, 1977), and
viduals may activate schemas and scripts that promote seeking when they are forced to evacuate, they tend to do so as a group
proximity to others (e.g., Mikulincer et al., 2002; Waters & (e.g., Cantril et al., 1940; Freeman & Cooper, 1940) or in fam-
Waters, 2006), even though this is sometimes not the safest ily units (Quarantelli & Dynes, 1977), thereby maintaining
strategy. Such proximity-seeking in cases of actual danger may proximity and contact with familiars. Bettelheim (1960)
have two disadvantages: slower identification of early signs of described how many Jewish families in Holland went into
danger and slower activation of defensive behavior. hiding as groups despite the fact that everyone knew they
Sime (1983, 1985) examined these disadvantages in a retro- would have a better chance of surviving as individuals.
spective study of reactions to a fire in a large coastal resort on In his study of the resort fire, Sime (1983, 1985) found that
the Isle of Man, Great Britain, in 1973. Shortly after a fire in 13 of the 50 people who died were in their groups when alerted
which 50 of 3,000 vacationers perished, accounts of 500 survi- about the fire, and the groups evidently delayed their depar-
vors were collected through interviews. These accounts suggest ture. In trying to escape in groups by whatever route they
that people who were physically closer to significant others chose, these people were caught by the encroaching smoke and
(e.g., family members) were less likely to react to ambiguous flames (Sime, 1983, p. 38). In fact, proximity-seeking beha-
cues of danger, such as noises and shouts, that occurred during vior seemed to increase the danger to family groups because
the early stages of the fire. They reacted only later, when unam- they were slower to escape and therefore increased the risk
biguous cues of danger, such as smoke, flames, and people run- of injury and death. Other disaster studies also show that people
ning while holding fire extinguishers, occurred. Subsequent in their groups are slow to leave disaster areas and instead wait
studies of survivors behavior during disasters also suggest that to be evacuated as a group (Aguirre et al., 1998; Fitzpatrick &
people who were together with familiar others were slow to Mileti 1991; Perry, 1994; Proulx, 2002, 2003).
react to danger (Aguirre, Wenger, & Vigo, 1998; Fitzpatrick Taken together, the evidence suggests that although
& Mileti, 1991; Perry, 1994; Proulx, 2002, 2003). In the pres- securely attached individuals are better at leading and coordi-
ence of significant others, many people seem to react slower to nating group activities, these advantages are partially offset
cues of imminent danger. by their slower identification of actual and imminent dangers
Research examining reactions to real or imagined dangers and their sometimes nonoptimal reactions to danger due to a
also provides indirect support for the hypothesis that securely wish to stay close to other people. In other words, securely
attached people react in nonoptimal ways to signs of danger. attached individuals action scripts, focused on seeking prox-
For example, in October 1938, Orson Welles broadcast a sec- imity to others in times of threat, may sometimes hamper their
tion from the science-fiction book The War of the Worlds survival and the survival of their group. This suggests that the
(Wells, 1898) as if it were genuine news. He dramatically survival chances of groups composed only of securely attached
depicted the advance of the Martians as a judgment day for people might in some important cases actually be lower than
humanity. Cantril, Gaudet, and Herzog (1940) studied the the survival chances of insecurely attached people.
famous broadcast and its outcomes. It appears that many of the
people who listened to the broadcast believed what they heard
and were frightened, but the vast majority did not flee. Instead
Insecure Attachment
they contacted relatives and friends in the city. Most of those Adult attachment research (e.g., Rom & Mikulincer, 2003) has
who fled did so only after other family members had been shown that relatively anxious or avoidant individuals often per-
brought together. In the devastating Southeast Michigan form less optimally in group activities than do relatively secure
Flint-Beecher tornado of June 1953, people also tended to turn ones. They may take the work less seriously, make fewer or
to and protect loved ones rather than flee from the threat (Form poorer quality contributions to a team, and have lower expec-
& Nosow, 1958). These findings are not surprising because the tations of contributing to the team effort. Nevertheless, in some
attachment systems main tendency and function is to seek dangerous situations these individuals may have advantages in
proximity as a way of attaining safety (Bowlby, 1969/1982). escaping or in helping their group members escape to safety.
But proximity seeking in cases of imminent disaster may not
be adaptive for individuals or their close relatives. The Adaptive Advantages of Attachment Anxiety: Mental
Bowlby (1973) noted that during and after disasters, no Schemas Related to Serving as a Sentinel. People who are
member of a family is content, or indeed able to do anything relatively high in attachment anxiety adopt hyperactivating
else, until all members of the family are gathered together strategies of affect regulation in times of threat or stress, which
(p. 91). Studies of behavior during fires also show that people heightens the monitoring of threat-related cues and results in
tend to converge and cluster (Bryan, 1985, 2002; Sime, 1983, exaggerated threat appraisal, almost regardless of the actual
1985). Governments and trained professionals have great dif- threat (Cassidy & Kobak, 1988; Mikulincer et al., 2000; Shaver
ficulty getting people to evacuate before and during disasters, & Mikulincer, 2002). Moreover, they react to threats by mini-
because traditional family ties often keep individual mem- mizing distance from others, signaling danger, and clinging
bers in the danger zone until it is too late (Hill & Hansen, (Feeney & Noller, 1990; Mikulincer et al., 1998). In this way,
1962, p. 217). these individuals may offset some of the deficiencies of

Downloaded from at Interdisciplinary Center for on May 4, 2010 131

132 Ein-Dor, Mikulincer, Doron, and Shaver

securely attached group members by reacting quickly and sentinels in other species contribute to the survival of their
strongly to early, perhaps ambiguous cues of actual imminent groups, (b) the early human environment was most probably
danger. For instance, in case of fire, anxiously attached individ- well supplied with circumstances and predators that threatened
uals may alert other group members to the first signs of danger humans, (c) in such an environment the benefits of sentinels
(e.g., ambiguous cues such as unusual noises, shuffling feet, or would have been great, and (d) anxiously attached group mem-
shouts). They may possess schemas and action tendencies that bers quick and intense reactions to threats were probably use-
make them good sentinels. That is, whenever they become sen- ful to the majority of group members who were more secure
sitive to a threat, they may call it to other group members and therefore less excitable. Their anxious schemas and action
attention sooner than they would attend to it on their own. tendencies, while possibly being troublesome at times, would
Many species of animals benefit from having sentinels in have nevertheless been beneficial at other times to group
their midst. For instance, birds (e.g., Miller, 2005; Platzen & members survival and eventual reproduction.
Magrath, 2005), rodents (e.g., Brudzynski, 2005; Sherman,
1980, 1981), various mammals (e.g., Fichtel, 2004), and
primates (e.g., Coss, Ramakrishnan, & Schank, 2005; Riede, The Adaptive Advantages of Avoidant Attachment. Indi-
Bronson, Hatzikirou, & Zuberbuhler, 2005) produce shrill viduals who are relatively avoidant with respect to attachment
alarm signals when they detect a potential threat. Moreover, are accustomed to looking out for their own interests and taking
since classic studies by Dunford (1977) and Sherman (1977) care of themselves, even if this sometimes occurs at others
demonstrated nepotistic benefits to ground squirrels from expense (Mikulincer & Shaver, 2007). This means that they are
issue alarm signals, other studies have found that animals likely to rely on self-protective fight-or-flight reactions in times
obtain both direct and indirect fitness benefits by risking their of danger. This defensive pattern has both disadvantages and
own safety (e.g., by calling a predators attention to them) advantages. In the face of danger, avoidant individuals may
while signaling their associates of danger (e.g., Blumstein, be primarily motivated to save themselves, but this tendency
Steinmetz, Armitage, & Daniel, 1997; Hoogland, 1995, 1996; may allow them to quickly discover a way to do so. Meanwhile,
Schwagmeyer, 1980). In other words, the existence of sentinels anxiously and securely attached individuals may focus much of
in a group often enhances the survival chances of group members. their attention on the whereabouts and welfare of close associ-
In similar ways, human group members can benefit from ates without focusing quickly and fully on how to escape.
anxiously attached individuals hyperactivating strategies. In Imagine an avoidant person in the presence of a dangerous
the course of evolution, humans lived in relatively small groups fire (of the kind we mentioned earlier). While taking quick pro-
or tribes of kin, so anxiously attached individuals might have tective action, the person may find an escape route or take
increased their own inclusive fitness by being sentinels while effective action to put out the fire or seal a door to keep the fire
also enhancing their groups overall contributions to the gene outside. Moreover, the avoidant person may be personally
pool. That is, heterogeneous groups that included some mem- effective and perhaps even ruthless if the situation calls for
bers who were anxiously attached might have been more suc- it; he or she is not overwhelmed by emotion when drastic action
cessful in terms of both survival and reproduction than might is required. Although there are obvious moral dangers in
groups composed entirely of secure individuals. By making behaving this way, there is little doubt that it can increase an
these arguments, we are not implying that anxiously attached avoidant persons survival chances while sometimes saving
individuals always behave as sentinels, whereas secure and other peoples lives, including the lives of group members
avoidant people do not. We suggest only that the probability about whom the avoidant individual may not care very deeply.
of identifying threats and calling other peoples attention to Saving lives is likely to contribute to the avoidant persons
them is higher for anxious individuals than for secure or avoi- inclusive fitness (keeping in mind that humans for most of their
dant people. existence have lived in groups of kin) and to the groups overall
The existence of this pattern could be due either to genes fitness.
that contribute to its formation or to genes that allow it to be Evidence for the influence of a few group members early
developed in response to environmental conditions. As the decisions to flee a dangerous situation can be found in the
developmental research literature on attachment shows that research literatures on military situations and natural disasters.
anxious attachment develops when primary attachment figures One of the most alarming sights for human beings is other peo-
are themselves anxious and unreliable (perhaps by being sensi- ple running from danger (e.g., Mawson, 1980). Marshall (1947)
tive to threats and oriented toward their own self-protection), it eloquently stated the following in writing about military
should be more common, as Belsky (1999) argued (although behavior during World War II:
for different reasons), when the social or physical environment
is conducive to parents being anxious and unreliable. This envi- It can be laid down as a general rule that nothing is more likely
ronment is likely to include conditions in which threats to collapse a line of infantry than the sight of a few of its num-
(whether from weather, predators, or human conflicts and ber in full and unexplained flight to the rear. . . . One or two or
wars) are prevalent but unpredictable. more men made a sudden run to the rear which others in the
In sum, anxious attachment and its relation to sentinel beha- vicinity did not understand. . . . In every case the testimony of
vior may have been adaptive at the group level, given that (a) all witnesses clearly [indicated] that those who started the

132 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 133

run . . . had a legitimate or at least a reasonable excuse for the behavior patterns may promote the inclusive fitness of group
action. (pp. 145146) members in its own way (see also Nettle, 2006). In an unpub-
lished and generally overlooked dissertation study of real-
It is also known that in dangerous situations people tend to world work teams, Kimmel (2003) found that teams containing
follow the route they see others taking (Mawson, 1980). only securely attached individuals were less effective in
Individuals who flee first (those, we suggest, who are likely accomplishing group tasks than were work teams containing
to be disproportionately avoidant) often clear a way by opening only 75% or 50% securely attached individuals, with the
emergency doors, breaking a window, or finding a safer place remaining team members having insecure attachment styles.
to hide. When their escape route is identified and cleared, oth-
ers can follow and take advantage of the escape route. Thus,
avoidant individuals may increase their own and their group
members chances of survival under emergency conditions. Preliminary Evidence Supporting SDT
As in the case of anxiously attached individuals acting as sen- Using SDTs basic postulates, we derived three key predictions
tinels, a group that contained at least some avoidant members that we have begun to test empirically. First, at the cognitive
might benefit from their presence even if the same individuals level, attachment-anxious and attachment-avoidant people will
at other times caused difficulties (e.g., engaging in mate be characterized by greater cognitive accessibility to sentinel
poaching or being irresponsible as parents; Schachner & and rapid fightflight scripts. Moreover, anxious individuals
Shaver, 2004). will have more rapid access to sentinel scripts (indicated by a
Also as in the case of anxious attachment, the existence of low threshold for detecting danger and a high propensity to
the avoidant attachment pattern could be due either to genes alert others), whereas avoidant individuals will be character-
that contribute directly to its formation or to genes that allow ized by greater accessibility of rapid fightflight scripts (indi-
it to develop under certain environmental conditions. As the cated by a high propensity to think immediately of ways to
developmental research literature on attachment shows that escape dangerous situations, mainly with the goal of saving
avoidant attachment develops when primary attachment figures themselves). Second, at the behavior level, anxious and avoi-
are themselves emotionally restricted, highly independent, and dant people will differ in the way they behave in threatening
somewhat selfish, it should be more common, as Belsky (1999) situations. Whereas the more anxious individuals will quickly
argued, when the social or physical environment makes selfish- identify threats and alert others to them, avoidant people will
ness and independence useful for survival. This environment is show rapid fight or flight responses without deliberating with
likely to include conditions in which altruism, including gener- other group members or waiting for their decisions. Third,
ous parental behavior, and interdependence interfere with sur- groups that are more heterogeneous with respect to attachment
vival. In a cross-cultural study of 55 countries, Schmitt et al. orientations will be more effective in dealing with threats (e.g.,
(2004) found that avoidant adult attachment is most common resolve emergency situations faster than less heterogeneous
in Africa, where mortality due to disease, famine, and war is groups).
exceptionally high, sometimes making warm, secure relations
with other people less rather than more conducive to survival.
In sum, diversity of attachment patterns in human groups, Cognitive Evidence
tribes, and societies may have been adaptively advantageous In a series of five studies, Ein-Dor, Mikulincer, and Shaver
over the span of human evolution. Although securely attached (2009b) have examined attachment-related variations in the
individuals may provide better care for their children and may cognitive accessibility of sentinel and rapid fightflight scripts.
be generally more effective than anxious or avoidant individu- In the first study, researchers asked participants to examine a
als at leading and coordinating groups (Davidovitz et al., 2007; picture of a small group of people in a threatening situation and
Rom & Mikulincer, 2003), they may also be slower than write a story about what would happen next. Judges blind to
anxious individuals to notice signs of impending danger and participants attachment scores received explanations about the
slower than avoidant individuals to show group members how sentinel and rapid fightflight scripts and were trained to code
to escape danger. each story according to specific and well-defined criteria (e.g.,
An implication of this analysis is that an extended family or noticing threat before others, warning others about the threat,
tribal group that is heterogeneous with respect to attachment acting without receiving help from others, reacting quickly
styles should be better, on average, at dealing with threats and without depending on others actions). As we predicted, parti-
survival problems than should homogenous groups. Heteroge- cipants scoring higher on attachment anxiety were more likely
neous groups should have the ability to detect potential prob- to generate stories that were congruent with the sentinel script.
lems and threats (with anxious group members acting as Likewise, participants scoring higher on avoidant attachment
sentinels); the ability to act quickly without much deliberation, were more likely to generate stories that included the main
negotiation, and compromise (with avoidant people acting as components of the rapid fightflight script. It is important to
models of quick self-protection); and the ability to manage note that these associations were statistically significant even
complex social tasks (with securely attached people acting as after we controlled for broad personality traits, socially desir-
stable leaders and coordinators of the group). Each of these able response biases, and verbal ability.

Downloaded from at Interdisciplinary Center for on May 4, 2010 133

134 Ein-Dor, Mikulincer, Doron, and Shaver

In the second and third studies, Ein-Dor et al. (2009b) cognitive framework for going beyond the schema-relevant
examined attachment-related variations in memory for core information given. Specifically, they found that having a
components of either the sentinel script (Study 2) or the rapid well-developed schema for a particular domain enables people
fightflight scripts (Study 3). Previous studies have shown that to generate more impressions of and conjectures about the
having a well-developed and highly accessible schema for a thoughts, feelings, intentions, and traits of a storys protago-
particular domain speeds up recognition of schema-relevant nistinformation that was not explicitly presented in the story.
information encountered in a previous learning task and We similarly hypothesized that anxious individuals well-
increases false recognition of schema-relevant information that developed sentinel script would allow them to process informa-
did not actually appear in the learning task. Both of these tion relevant to the sentinel script in a deeper fashion, whereas
patterns are indications of schema-biased information process- avoidant individuals well-developed rapid fightflight script
ing (e.g., Lurigio & Carroll, 1985; Markus, 1977; Roediger & would allow them to process information relevant to this script
McDermott, 1995). We hypothesized that having a well- in a deeper fashion.
developed and highly accessible sentinel script would cause Participants were asked to read a story that included the
attachment-anxious people to quickly recognize real script- major components of either the sentinel script (Study 4) or the
relevant information and produce more false memories for rapid fightflight script (Study 5). They then generated recol-
items that might have occurred in the learning task but actually lections (actual facts presented in the story) and impressions
did not. Similarly, having a well-developed and highly accessi- (inferences, feelings, and opinions) about the story. Two inde-
ble rapid fightflight script would cause avoidant people to pendent judges placed each participants responses into one of
quickly recognize and falsely recall information relevant to that the following categories: script-relevant recollections, script-
script. relevant impressions, neutral (script-irrelevant) recollections,
Participants were shown a video clip of a young woman and neutral impressions. The results indicated that attachment
answering threat-relevant and threat-irrelevant questions. They anxiety, but not avoidance, was associated with generating
then performed a recognition task related to her answers. For more inferences concerning the sentinel script. However, it did
half of the participants, the target womans answers to threat- not explain individual differences in the number of sentinel-
relevant questions were congruent with either the sentinel script recollections or the number of inferences made about
script (Study 2) or the rapid fightflight script (Study 3). For script-irrelevant statements. Similarly, as predicted, avoidant
the remaining participants, the target womans answers were attachment was associated with generating more inferences
incongruent with both of these scripts. In the recognition task, about the rapid fightflight script. However, it did not explain
participants received statements that had appeared among the individual differences in the number of rapid fightflight recol-
target womans answers (old items) and statements that lections or the number of recollections or inferences made
resembled the target womans answers semantically but had about script-irrelevant issues. In addition, attachment anxiety
not actually been among her answers (new items). In this task, was found to impair recall of information congruent with the
participants were asked to decide whether or not each of these rapid fightflight script.
statements had actually appeared in the video clip. For each Overall, Ein-Dor et al.s (2009b) findings suggest that
participant and each sentence category (neutral, threat), we anxiously attached people possess highly accessible and well-
calculated two scores: the percentage of trials in which old sen- organized implicit knowledge about a sequence of events that
tences were accurately recognized (hits), and the percentage goes from monitoring and quickly reacting to potential sources
of trials in which new sentences were incorrectly believed to of threats to alerting others about the imminent danger and
have appeared in the video clip (false memories). We also minimizing distance from others. This highly accessible
calculated average reaction times (RTs) for hits and false knowledge structure probably contributes to the well-
memories. documented tendency of anxiously attached individuals to
In line with predictions, attachment anxiety, but not avoid- become highly distressed in face of threats and to cope with all
ance, was significantly associated with more schema-biased kinds of difficulties by catastrophizing, directing attention to
memories and faster recognition of information that was con- threat-related information, vigorously expressing needs and
gruent with the sentinel script. Similarly, as predicted, avoidant vulnerabilities, and desperately seeking proximity to others and
attachment, but not anxious attachment, was significantly asso- support and comfort from them (see Mikulincer & Shaver,
ciated with more schema-biased memory and faster recognition 2007, for a review).
of information that was congruent with the rapid fightflight The findings also suggest that avoidant people possess highly
script. When the information in the video clip was not congru- accessible and well-organized implicit knowledge about a
ent with the rapid fightflight script, more avoidant participants sequence of behaviors beginning with rapid efforts to preserve
reacted with less schema-biased memory. oneselffight or flight responseswithout deliberating or coor-
In the fourth and fifth studies, Ein-Dor et al. (2009b) exam- dinating responses with other people or expecting help from
ined attachment-related variations in the processing of them. This highly accessible rapid fightflight script may under-
information relevant to either the sentinel script (Study 4) or the lie research findings showing that avoidant individuals are reluc-
rapid fightflight script (Study 5). Markus, Smith, and More- tant to seek support in times of trouble, keep somewhat distant
land (1985) showed that a well-developed schema provides a from and independent of other people, suppress distress-related

134 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 135

thoughts, and emphasize autonomy and self-efficacy (see social processes in adult life. One consistent finding has been
Mikulincer & Shaver, 2007, for a review). that securely attached individuals are happier and better
adapted in virtually every area of daily life. This made us
Behavioral Evidence wonder why approximately half of the human population is
insecure with respect to attachment. It seemed to be an evolu-
In an initial test of SDTs behavioral predictions concerning tionary paradox. SDT was devised to explain this paradox. It is
behavior in groups, Ein-Dor, Mikulincer, and Shaver (2009a) based on the possibility that attachment patterns that seem trou-
asked Israeli undergraduates to complete the ECR scale at the blesome and dysfunctional at an individual or dyadic level
beginning of the semester and then assessed their actual beha- make sense when considered at the level of groups, tribes, and
vior during an experimentally induced threatening situation in a societies. At that higher level of analysis, it seems possible that
small-group laboratory situation. Participants were invited to attachment-style heterogeneity is conducive to survival when a
the laboratory in groups of three (46 groups in all). Upon arri- group encounters life-threatening dangers. SDT makes empiri-
val, an experimenter took the 3 participants to a large room, sat cally testable predictions, and we have already found in the two
them beside a long table, and asked them to complete a battery preliminary studies briefly summarized here that two of SDTs
of questionnaires while he prepared the computers for the predictions are supported by experimental data. It therefore
experiment. The experimenter then exited the room and closed seems worthwhile to test the theory further.
the door behind him. Participants were filmed by hidden cam- If SDT continues to receive empirical support, it will be
eras throughout the session. Ten feet behind the participants table important to determine the best explanation from an evolution-
was another table on which an SVGA monitor displayed a generic ary viewpoint. If a mixture of attachment patterns is useful at the
desktop graphic; on the floor nearby was an apparently attached group level, this suggests one or both of two alternative evolu-
PC, which was actually a nontoxic party smoke machine. Exactly tionary accounts. First, it might imply that different attachment
1 minute after the experimenter departed, he began sending patterns are related to different genetic alleles (as already
smoke into the room through the bogus PC computer, making it indicated by Crawford et al., 2007; Donnellan et al., 2008; and
seem to participants that the computer had caught on fire. The Gillath et al., 2008), in which case inclusive fitness or group-
experiment ended when the participants either exited the room level selection processes might have determined the relative fre-
or tried to deal with the smoking computer. quencies of the different attachment patterns in the human pop-
Two judges, blind to participants attachment scores and the ulation. To date, however, there is little evidence for strong,
SDT hypotheses, recorded the following information: (a) the direct genetic determination of attachment patterns and some
identity of the participant who was the first to detect the pres- evidence against it (see Vaughn, Bost, & van IJzendoorn,
ence of the smoke in the room; (b) the identity of the participant 2008, for a review of genetic studies of infant attachment pat-
who was the first to react, either by exiting the room or attempt- terns). The second possibility is that evolution generated a
ing to deal with the danger (whichever came first); (c) the capacity for facultative development of attachment patterns in
amount of time (in seconds) from turning on the smoke response to environmental pressures. This is the kind of explana-
machine to participants detection of smoke in their room; and tion offered by Belsky (1999), although he based his version of
(d) the amount of time (in seconds) from the onset of the smoke the explanation on mating processes rather than survival.
to the conclusion of the session. In addition, the judges rated In either case, it will be important to discern why the relative
the effectiveness of each group in dealing with the situation frequencies of the major attachment patterns seem to be similar
on a scale ranging from 1 (not at all) to 7 (very much). around the world. If there really is such a general similarity, it
Consistent with SDT predictions, the probability of identify- suggests either similarities in the average kinds and levels of
ing the presence of smoke in the room was significantly higher threats to groups or frequency-based selection (West-Eberhard,
for individuals who scored relatively high on attachment anxiety. 2003). An argument against this proportional constancy is that
In addition, the probability of reacting first to the danger was sig- insecure attachment is more common in lower socioeconomic
nificantly higher for individuals who scored relatively high on groups (Steele & Steele, 2008) and, as already mentioned, in
avoidant attachment. These associations were still significant environmentally challenging parts of Africa (Schmitt et al.,
after controlling for participants levels of extroversion and neu- 2004). These considerations suggest that if SDT is supported
roticism. Moreover, more heterogeneous groups in terms of in future studies, it needs to be developed more fully with respect
attachment orientations were rated by judges to be more effec- to alternative biological and evolutionary explanations.
tive in dealing with the dangerous situation and took less time
to detect and deal with the danger. Overall, the findings lend Declaration of Conflicting Interests
impressive initial support for SDT hypotheseshypotheses that The authors declared that they had no conflicts of interest with respect
would not have been proposed or tested without SDT. to their authorship or the publication of this article.

Concluding Comments Aguirre, B.E., Wenger, D., & Vigo, G. (1998). A test of the emergent
Over the past 20 years, numerous studies have demonstrated norm theory of collective behavior. Sociological Forum, 13,
the relevance of attachment theory to many psychological and 301311.

Downloaded from at Interdisciplinary Center for on May 4, 2010 135

136 Ein-Dor, Mikulincer, Doron, and Shaver

Ainsworth, M.D.S., Blehar, M.C., Waters, E., & Wall, S. (1978). Brennan, K.A., Shaver, P.R., & Tobey, A.E. (1991). Attachment
Patterns of attachment: Assessed in the strange situation and at styles, gender, and parental problem drinking. Journal of Social
home. Hillsdale, NJ: Erlbaum. and Personal Relationships, 8, 451466.
Alexander, R.D. (1987). The biology of moral systems. New York: Bretherton, I., & Munholland, K.A. (2008). Internal working models
Aldine de Gruyter. in attachment relationships: Elaborating a central construct in
Axelrod, R. (1984). The evolution of cooperation. New York: Basic attachment theory. In J. Cassidy & P. Shaver (Eds.), Handbook
Books. of attachment: Theory, research and clinical application (2nd rev.
Baker, G.W., & Chapman, D.W. (1962). Man and society in disaster. ed.), New York: Guilford Press.
New York: Basic Books. Brewer, M.B., & Caporael, L.R. (1990). Selfish genes vs. selfish peo-
Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2004). No ple: Sociobiology as origin myth. Motivation Emotion. 14,
association of the dopamine D4 receptor (DRD4) and -521 C/T 237243.
promoter polymorphisms with infant attachment disorganization. Brudzynski, S.M. (2005). Principles of rat communication: Quantita-
Attachment and Human Development, 6, 211218. tive parameters of ultrasonic calls in rats. Behavior Genetics, 35,
Bartholomew, K. (1990). Avoidance of intimacy: An attachment 8592.
perspective. Journal of Social and Personal Relationships, 7, Brussoni, M.J., Jang, K.L., Livesley, W., & MacBeth, T.M. (2000).
147178. Genetic and environmental influences on adult attachment styles.
Bartholomew, K., & Horowitz, L.M. (1991). Attachment styles among Personal Relationships, 7, 283289.
young adults: A test of a four-category model. Journal of Person- Bryan, J.L. (1985). Convergence clusters: A phenomenon of human
ality and Social Psychology, 61, 226244. behavior seen in selected high-rise building fires. Fire Journal,
Beller, E.K., & Pohl, A. (1986, April). The strange situation revisited. 79(6), 2730.
Paper presented at the International Conference on Infant Studies, Bryan, J.L. (2002). A selected historical review of human behavior in
Beverly Hills, CA. fire. Fire Protection Journal, 16, 410.
Belsky, J. (1997). Theory testing, effect-size evaluation, and differen- Buss, D.M. (1988a). The evolution of human intrasexual competition:
tial susceptibility to rearing influence: The case of mothering and Tactics of mate attraction. Journal of Personality and Social Psy-
attachment. Child Development, 68, 598600. chology, 54, 616628.
Belsky, J. (1999). Modern evolutionary theory and patterns of attach- Buss, D.M. (1988b). From vigilance to violence: Tactics of mate
ment. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment: retention in American undergraduates. Ethology and Sociobiology,
Theory, research, and clinical applications (pp. 141161). New 9, 291317.
York: Guilford Press. Buss, D.M. (1989). Sex differences in human mate preferences: Evo-
Belsky, J., Steinberg, L., & Draper, P. (1991). Childhood experience, lutionary hypotheses tested in 37 cultures. Behavioral and Brain
interpersonal development, and reproductive strategy: An evolu- Sciences, 12, 149.
tionary theory of socialization. Child Development, 62, 647670. Buss, D.M. (1995). Evolutionary psychology: A new paradigm for
Benoit, D., & Parker, K.C.H. (1994). Stability and transmission of psychological science. Psychological Inquiry, 6, 130.
attachment across three generations. Child Development, 65, Buss, D.M. (1999). Evolutionary psychology: The new science of the
14441456. mind. Boston: Allyn and Bacon.
Bettelheim, B. (1960). The informed heart: Autonomy in a mass age. Buss, D.M., Larsen, R.J., Westen, D., & Semmelroth, J. (1992). Sex
New York: Free Press. differences in jealousy: Evolution, physiology, and psychology.
Birnbaum, G.E., Orr, I., Mikulincer, M., & Florian, V. (1997). When Psychological Science, 3, 251255.
marriage breaks up: Does attachment style contribute to coping and Cantril, H., Gaudet, H., & Herzog, H. (1940). The invasion from Mars.
mental health? Journal of Social and Personal Relationships, 14, Princeton, NJ: Princeton University Press.
643654. Carnelley, K.B., & Janoff-Bulman, R. (1992). Optimism about love
Blumstein, D.T., Steinmetz, J., Armitage, K.B., & Daniel, J.C. (1997). relationships: General vs. specific lessons from ones personal
Alarm calling in yellow-bellied marmots: II. The importance of experiences. Journal of Social and Personal Relationships, 9, 520.
direct fitness. Animal Behavior, 53, 173184. Carnelley, K.B., Pietromonaco, P.R., & Jaffe, K. (1994). Depression,
Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation: Anxiety working models of others, and relationship functioning. Journal of
and anger. New York: Basic Books. Personality and Social Psychology, 66, 127140.
Bowlby, J. (1980). Attachment and loss: Vol. 3. Sadness and depres- Carpenter, B.D. (2001). Attachment bonds between adult daughters
sion. New York: Basic Books. and their older mothers: Associations with contemporary caregiv-
Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment, (2nd ed.). ing. Journals of Gerontology: Series B. Psychological Sciences
New York: Basic Books. (Original work published 1969) and Social Sciences, 56B, 257266.
Bowlby, J. (1988). A secure base: Clinical applications of attachment Carpenter, S.L. (1988). Self-relevant and goal-directed processing in
theory. London: Routledge. the recall and weighting of information about others. Journal of
Brennan, K.A., Clark, C.L., & Shaver, P.R. (1998). Self-report mea- Experimental Social Psychology, 24, 310332.
surement of adult romantic attachment: An integrative overview. Cassidy, J., & Berlin, L.J. (1994). The insecure/ambivalent pattern of
In J.A. Simpson & W.S. Rholes (Eds.), Attachment theory and attachment: Theory and research. Child Development, 65,
close relationships (pp. 4676). New York: Guilford Press. 971991.

136 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 137

Cassidy, J., & Kobak, R.R. (1988). Avoidance and its relationship Dunford, C. (1977). Kin selection for ground squirrel alarm calls.
with other defensive processes. In J. Belsky & T. Nezworski American Naturalist, 58, 782785.
(Eds.), Clinical implications of attachment (pp. 300323). Eberly, M.B., & Montemayor, R. (1998). Doing good deeds: An
Hillsdale, NJ: Erlbaum. examination of adolescent prosocial behavior in the context of
Cassidy, J., & Shaver, P.R. (Eds.). (1999). Handbook of attachment: parent-adolescent relationships. Journal of Adolescent Research,
Theory, research, and clinical applications. New York: Guilford 13, 403432.
Press. Eberly, M.B., & Montemayor, R. (1999). Adolescent affection and
Cassidy, J., & Shaver, P.R. (Eds.). (2008). Handbook of attachment: helpfulness toward parents: A 2-year follow-up. Journal of Early
Theory, research, and clinical applications, (2nd ed.). New Adolescence, 19, 226248.
York: Guilford Press. Ein-Dor, T., Mikulincer, M., & Shaver, P. (2009a). Attachment
Chisholm, J.S. (1996). The evolutionary ecology of attachment orga- insecurities and the behavioral reactions to an experimentally-
nization. Human Nature, 7, 138. induced threat. Manuscript submitted for publication.
Cicirelli, V.G. (1993). Attachment and obligation as daughters Ein-Dor, T., Mikulincer, M., & Shaver, P. (2009b). Attachment inse-
motives for caregiving behavior and subsequent effect on curities and the processing of threat-related information: Studying
subjective burden. Psychology and Aging, 8, 144155. the scripts involved in insecure peoples coping strategies. Manu-
Collins, N.L., & Read, S.J. (1990). Adult attachment, working models, script submitted for publication.
and relationship quality in dating couples. Journal of Personality Feeney, J.A. (2004). Transfer of attachment from parents to romantic
and Social Psychology, 58, 644663. partners: Effects of individual and relationship variables. Journal
Cosmides, L. (1989). The logic of social exchange: Has natural of Family Studies, 10, 220238.
selection shaped how humans reason? Studies with the Wason Feeney, J. A., & Noller, P. (1990). Attachment style as a predictor of
selection task. Cognition, 31, 187276. adult romantic relationships. Journal of Personality and Social
Coss, R.G., Ramakrishnan, U., & Schank, J. (2005). Recognition of Psychology, 58, 281291.
partially concealed leopards by wild bonnet macaques (Macaca Fiala, K.B. (1991). Attachment and psychosocial functioning of
radiata) the role of the spotted coat. Behavioural Processes, 68, depressed, remitted depressed, and nondepressed women and their
145163. partners. Unpublished doctoral dissertation, University of Massa-
Crawford, T.N., Livesley, W.J., Jang, K.L., Shaver, P.R., Cohen, P., & chusetts, Amherst.
Ganiban, J. (2007). Insecure attachment and personality disorder: Fichtel, C. (2004). Reciprocal recognition in sifaka (Propithecus ver-
A twin study of adults. European Journal of Personality, 21, reauxi verreauxi) and redfronted lemur (Eulemur fulvus rufus)
191208. alarm calls. Animal Cognition, 7, 4552.
Crispi, E.L., Schiaffino, K., & Berman, W.H. (1997). The contribution Finkel, D., & Matheny, A.P. (2000). Genetic and environmental influ-
of attachment to burden in adult children of institutionalized ences on a measure of infant attachment security. Twin Research,
parents with dementia. Gerontologist, 37, 5260. 3, 242250.
Cross, S.E., & Markus, H.R. (1994). Self-schemas, possible selves, Fitzpatrick, C., & Mileti, D.S. (1991). Motivating public evacuation.
and competent performance. Journal of Educational Psychology, International Journal of Mass Emergencies and Disasters, 9,
86, 423438. 137152.
Daly, M., Wilson, M.I., & Weghorst, S.J. (1982). Male sexual jea- Flinn, M.V. (1988). Step- and genetic parent/offspring relationships in
lousy. Ethology and Sociobiology, 3, 1127. a Caribbean village. Ethology and Sociobiology, 9, 335369.
Davidovitz, R., Mikulincer, M., Shaver, P.R., Ijzak, R., & Popper, M. Fonagy, P., Steele, H., & Steele, M. (1991). Maternal representations
(2007). Leaders as attachment figures: Their attachment of attachment during pregnancy predict the organization of infant-
orientations predict leadership-related mental representations and mother attachment at one year of age. Child Development, 62,
followers performance and mental health. Journal of Personality 891905.
and Social Psychology, 93, 632650. Form, W.H., & Nosow, S. (1958). Community in disaster. New York:
Dawkins, R. (1994). Burying the vehicle. Behavioral and Brain Harper.
Sciences, 17, 616617. Fraley, R.C., & Shaver, P.R. (1997). Adult attachment and the sup-
Del-Giudice, M. (2009). Sex, attachment, and the development of pression of unwanted thoughts. Journal of Personality and Social
reproductive strategies. Behavioral and Brain Sciences, 32, 167. Psychology, 73, 10801091.
Donnellan, M.B., Burt, S.A., Levendosky, A.A., & Klump, K.L. Fraley, R.C., & Waller, N.G. (1998). Adult attachment patterns: A test
(2008). Genes, personality, and attachment in adults: A multivari- of the typological model. In J.A. Simpson & W.S. Rholes (Eds.),
ate behavioral genetic analysis. Personality and Social Psychology Attachment theory and close relationships (pp. 77114). New
Bulletin, 34, 316. York: Guilford Press.
Doron, G., & Kyrios, M. (2005). Obsessive compulsive disorder: A Freeman, C.D., & Cooper, D. (1940). The road to Bordeaux. New
review of possible specific internal representations within a broader York: Cresset Press.
cognitive theory. Clinical Psychology Review, 25, 415432. Friedberg, N.L., & Lyddon, W.J. (1996). Self-other working models and
Dozier, M., & Kobak, R. (1992). Psychophysiology in attachment eating disorders. Journal of Cognitive Psychotherapy, 10, 193203.
interviews: Converging evidence for deactivating strategies. Child Gazzaniga, M.S. (2008). Human: The science behind what makes us
Development, 63, 14731480. unique. New York: HarperCollins.

Downloaded from at Interdisciplinary Center for on May 4, 2010 137

138 Ein-Dor, Mikulincer, Doron, and Shaver

Gillath, O., Shaver, P.R., Baek, J.-M., & Chun, D.S. (2008). Genetic Kinston, W., & Rosser, R. (1974). Disasters: Effects on mental and
correlates of adult attachment style. Personality and Social physical health. Journal of Psychosomatic Research, 18, 437456.
Psychology Bulletin, 34, 13961405. Kirkpatrick, L.A., & Hazan, C. (1994). Attachment styles and close
Grossmann, K.E., Grossmann, K., Huber, F., & Wartner, U. (1981). relationships: A four-year prospective study. Personal Relation-
German childrens behavior toward their mothers at 12 months and ships, 1, 123142.
their fathers at 18 months in Ainsworths Strange Situation. Kirkpatrick, L.A., & Shaver, P.R. (1990). Attachment theory and reli-
International Journal of Behavioral Development, 7, 157181. gion: Childhood attachments, religious beliefs, and conversion.
Grossmann, K.E., Grossmann, K., & Waters, E. (Eds.). (2005). Journal for the Scientific Study of Religion, 29, 315334.
Attachment from infancy to adulthood. New York: Guilford Press. Kobak, R., Cole, H., Ferenz-Gillies, R., & Fleming, W. (1993). Attach-
Hamilton, W.D. (1964). The genetical evolution of social behaviour: I ment and emotional regulation during mother-teen problem solving:
and II. Journal of Theoretical Biology, 7, 152. A control theory analysis. Child Development, 64, 231245.
Hamilton, W.D. (1975). Innate social aptitudes in man: An approach Kobak, R., & Sceery, A. (1988). Attachment in late adolescence:
from evolutionary genetics. In R. Fox (Ed.), Biosocial anthropol- Working models, affect regulation, and representations of self and
ogy (pp. 133155). London: Malaby Press. others. Child Development, 59, 135146.
Harris, C.R. (2003). A review of sex differences in sexual jealousy, Kunce, L.J., & Shaver, P.R. (1994). An attachment-theoretical
including self-report data, psychophysiological responses, inter- approach to caregiving in romantic relationships. In
personal violence, and morbid jealousy. Personality and Social K. Bartholomew & D. Perlman (Eds.), Attachment processes in
Psychology Review, 7, 102128. adulthood (Advances in personal relationships (Vol. 5, pp. 205
Hazan, C., & Hutt, M. (1993). Patterns of adaptations: Attachment 237). London: Jessica Kingsley.
differences in psychosocial functioning during the first year of col- Lamb, M.E., Hwang, C.P., Frodi, M., & Frodi, M. (1982). Security of
lege. Unpublished manuscript, Cornell University, Ithaca, NY. mother and fatherinfant attachment and its relation to sociability
Hazan, C., & Shaver, P.R. (1987). Romantic love conceptualized as an with strangers in traditional and non-traditional Swedish families.
attachment process. Journal of Personality and Social Psychology, Infant Behavior and Development, 5, 355367.
52, 511524. Lurigio, A.J., & Carroll, J.S. (1985). Probation officers schemata of
Hazan, C., & Shaver, P.R. (1990). Love and work: An attachment- offenders: Content, development, and impact on treatment deci-
theoretical perspective. Journal of Personality and Social Psychol- sions. Journal of Personality and Social Psychology, 48,
ogy, 59, 270280. 11121126.
Hazan, C., & Zeifman, D. (1999). Pair-bonds as attachments: Lussier, Y., Sabourin, S., & Turgeon, C. (1997). Coping strategies as
Evaluating the evidence. In J. Cassidy & P.R. Shaver (Eds.), Hand- moderators of the relationship between attachment and marital
book of attachment: Theory, research, and clinical applications adjustment. Journal of Social and Personal Relationships, 14,
(pp. 336354). New York: Guilford. 777791.
Henderson, S. (1977). The social network, support and neurosis: The Main, M. (1990). Cross-cultural studies of attachment organization:
function of attachment in adult life. British Journal of Psychiatry, Recent studies, changing methodologies, and the concept of
131, 185191. conditional strategies. Human Development, 33, 4861.
Hesse, E. (2008). The Adult Attachment Interview: Historical and cur- Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy,
rent perspectives. In J. Cassidy & P.R. Shaver (Eds.), Handbook of childhood, and adulthood: A move to the level of representation.
attachment: Theory, research, and clinical applications (2nd ed., Monographs of the Society for Research in Child Development,
pp. 552598). New York: Guilford Press. 50, 66104.
Hill, R., & Hansen, D.A. (1962). Families in disaster. In G.W. Baker & Marks, I.M. (1987). Fears, phobias, and rituals. New York: Oxford
D.W. Chapman (Eds.), Man and society in disaster (pp. 185221). University Press.
New York: Basic Books. Markus, H. (1977). Self-schemata and processing information about
Hoogland, J. (1995). The black-tailed prairie dog: Social life of a bur- the self. Journal of Personality and Social Psychology, 35, 6378.
rowing mammal. Chicago: University of Chicago Press. Markus, H. (1980). The self in thought and memory. In D.M. Wegner
Hoogland, J. (1996). Why do Gunnisons prairie dogs give anti- & R.R. Vallacher (Eds.), The self in social psychology (pp. 102
predator calls? Animal Behavior, 51, 871880. 130). New York: Oxford University Press.
Inglehart, M.R., Markus, H., & Brown, D.R. (1989). The effects of Markus, H. (1983). Self-knowledge: An expanded view. Journal of
possible selves on academic achievement: A panel study. In Personality, 51, 543565.
J.P. Forgas & J.M. Innes (Eds.), Recent advances in social psychol- Markus, H., Crane, M., Bernstein, S., & Siladi, M. (1982). Self-
ogy: An international perspective (pp. 469477). Amsterdam: schemas and gender. Journal of Personality and Social Psychol-
Elsevier Science Publishers B. V. (North-Holland). ogy, 42, 3850.
Kanagawa, C., Cross, S.E., & Markus, H. R. (2001). Who am I? Markus, H., Cross, S., & Wurf, E. (1990). The role of the self-system
The cultural psychology of the conceptual self. Personality and in competence. In R.J. Sternberg & J. Kolligian, Jr. (Eds.), Compe-
Social Psychology Bulletin, 27, 90103. tence considered (pp. 205225). New Haven, CT: Yale University
Kimmel, M. (2003). Attachment patterns and social support processes Press.
in work teams. Unpublished doctoral dissertation, Technion-Israel Markus, H., & Nurius, P. (1986). Possible selves. American Psychol-
Institute of Technology, Haifa, Israel. ogist, 41, 954969.

138 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 139

Markus, H., & Ruvolo, A. (1989). Possible selves: Personalized Mikulincer, M., & Orbach, I. (1995). Attachment styles and repressive
representations of goals. In L. Pervin (Ed.), Goal concepts in person- defensiveness: The accessibility and architecture of affective
ality and social psychology (pp. 211241). Hillsdale, NJ: Erlbaum. memories. Journal of Personality and Social Psychology, 68,
Markus, H.R., & Sentis, K.P. (1982). The self in social information 917925.
processing. In J. Suls (Ed.), Psychological perspectives on the self Mikulincer, M., Orbach, I., & Iavnieli, D. (1998). Adult attachment
(Vol. 1, pp. 4170). Erlbaum, Hillsdale, NJ. style and affect regulation: Strategic variations in subjective self-
Markus, H., Smith, J., & Moreland, R.L. (1985). The role of the self- other similarity. Journal of Personality and Social Psychology,
concept in the perception of others. Journal of Personality and 75, 436448.
Social Psychology, 49, 14941512. Mikulincer, M., & Shaver, P.R. (2003). The attachment behavioral
Markus, H., & Wurf, E. (1987). The dynamic self-concept: A social system in adulthood: Activation, psychodynamics, and interperso-
psychological perspective. In M.R. Rosenweig & L.W. Porter nal processes. In M.P. Zanna (Ed.), Advances in experimental
(Eds.), Annual review of psychology (Vol. 38, pp. 299337). social psychology (pp. 53152). New York: Academic Press.
Palo Alto, CA: Annual Reviews. Mikulincer, M., & Shaver, P.R. (2005). Attachment security, compas-
Marshall, S.L.A. (1947). Men against fire. New York: Morrow. sion, and altruism. Current Directions in Psychological Science,
Mawson, A.R. (1978, August). Panic behavior: A review and a new 14, 3438.
hypothesis. Paper presented at the 9th World Congress of Mikulincer, M., & Shaver, P.R. (2007). Attachment in adulthood:
Sociology, Uppsala, Sweden. Structure, dynamics, and change. New York: Guilford Press.
Mawson, A.R. (1980). Is the concept of panic useful for study pur- Miller, G. (2005). Behavior: Bird alarm calls size up predators. Sci-
poses? In B. Levin (Ed.), Behavior in fires [NBS Report ence, 308, 18531855.
NBSIR-802070]. Washington, DC: U.S. Department of Miyake, K., Chen, S.J., & Campos, J.J. (1985). Infant temperament,
Commerce. mothers mode of interaction, and attachment in Japan: An interim
Mawson, A.R. (2005). Understanding mass panic and other collective report. Monographs of the Society for Research in Child Develop-
responses to threat and disaster. Psychiatry: Interpersonal & ment, 50(12, Serial no. 209), 276297.
Biological Processes, 68, 95113. Nettle, D. (2006). The evolution of personality variation in humans
Mickelson, K.D., Kessler, R.C., & Shaver, P.R. (1997). Adult attach- and other animals. American Psychologist, 61, 622631.
ment in a nationally representative sample. Journal of Personality OConnor, T.G., & Croft, C.M. (2001). A twin study of attachment in
and Social Psychology, 73, 10921106. preschool children. Child Development, 72, 15011511.
Mikulincer, M. (1998). Adult attachment style and affect regulation: Oyserman, D., & Markus, H. (1990). Possible selves and delinquency.
Strategic variations in self-appraisals. Journal of Personality and Journal of Personality and Social Psychology, 59, 112125.
Social Psychology, 75, 420435. Perry, R.W. (1994). A model of evacuation compliance behavior. In
Mikulincer, M., Birnbaum, G., Woddis, D., & Nachmias, O. (2000). R.R. Dynes & K.J. Tierney (Eds.), Disasters, collective behavior,
Stress and accessibility of proximity-related thoughts: Exploring and social organization (pp. 8598). Newark: University of
the normative and intraindividual components of attachment the- Delaware Press.
ory. Journal of Personality and Social Psychology, 78, 509523. Platzen, D., & Magrath, R.D. (2005). Adaptive differences in response
Mikulincer, M., & Florian, V. (1995). Appraisal of and coping with a to two types of parental alarm calls in altricial nestlings. Proceed-
real-life stressful situation: The contribution of attachment styles. ings of the Royal Society of London, Series B, 272, 11011106.
Personality and Social Psychology Bulletin, 21, 406414. Proulx, G. (2002, April 1516). Understanding human behavior in
Mikulincer, M., & Florian, V. (1998). The relationship between adult stressful situations. Paper presented in a workshop at the National
attachment styles and emotional and cognitive reactions to Academy of Sciences, Delegate Binder Section 7, Washington,
stressful events. In J.A. Simpson & W.S. Rholes (Eds.), Attach- DC.
ment theory and close relationships (pp. 143165). New York: Proulx, G. (2003). Researchers learn from World Trade Center survi-
Guilford Press. vors accounts. Construction Innovation, 8, 13.
Mikulincer, M., Florian, V., & Tolmacz, R. (1990). Attachment styles Quarantelli, E.L., & Dynes, R. (1977). Response to social crisis and
and fear of personal death: A case study of affect regulation. Jour- disaster. Annual Review of Sociology, 3, 2349.
nal of Personality and Social Psychology, 58, 273280. Riede, T., Bronson, E., Hatzikirou, B., & Zuberbuhler, K. (2005). The
Mikulincer, M., Florian, V., & Weller, A. (1993). Attachment styles, production mechanisms of Diana monkey alarm calls: Morpholo-
coping strategies, and posttraumatic psychological distress: The gical data and a model. Journal of Human Evolution, 48, 8596.
impact of the Gulf War in Israel. Journal of Personality and Social Riskind, J.H., Williams, N.L., Altman, M.D., Black, D.O.,
Psychology, 64, 817826. Balaban, M.S., & Gessner, T.L. (2004). Developmental antece-
Mikulincer, M., Gillath, O., & Shaver, P.R. (2002). Activation of dents of the looming maladaptive style: Parental bonding and par-
the attachment system in adulthood: Threat-related primes increase ental attachment insecurity. Journal of Cognitive Psychotherapy,
the accessibility of mental representations of attachment figures. 18, 4352.
Journal of Personality and Social Psychology, 83, 881895. Roediger, H.L., III., & McDermott, K.B. (1995). Creating false mem-
Mikulincer, M., & Nachshon, O. (1991). Attachment styles and ories: Remembering words not presented in lists. Journal of
patterns of self-disclosure. Journal of Personality and Social Experimental Psychology: Learning, Memory, and Cognition,
Psychology, 61, 321331. 21, 803814.

Downloaded from at Interdisciplinary Center for on May 4, 2010 139

140 Ein-Dor, Mikulincer, Doron, and Shaver

Rom, E., & Mikulincer, M. (2003). Attachment theory and group Smith, E.R., Murphy, J., & Coats, S. (1999). Attachment to groups:
processes: The association between attachment style and group- Theory and management. Journal of Personality and Social Psy-
related representations, goals, memories, and functioning. Journal chology, 77, 94110.
of Personality and Social Psychology, 84, 12201235. Sober, E., & Wilson, D.S. (1998). Unto others: The evolution and psy-
Rothbard, J.J., & Shaver, P.R. (1994). Continuity of attachment across chology of unselfish behavior. Cambridge, MA: Harvard
the life span. In M.B. Sperling & W.H. Berman (Eds.), Attachment University Press.
in adults (pp. 3171). New York: Guilford Press. Sroufe, L.A., & Waters, E. (1977). Attachment as an organizational
Safford, S.M., Alloy, L.B., Crossfield, A.G., Morocco, A.M., & construct. Child Development, 48, 11841199.
Wang, J.C. (2004). The relationship of cognitive style and attach- Steele, H., & Steele, M. (Eds.). (2008). Clinical applications of the
ment style to depression and anxiety in young adults. Journal of Adult Attachment Interview. New York: Guilford Press.
Cognitive Psychotherapy, 18, 2541. Symons, D. (1979). The evolution of human sexuality. New York:
Sagi, A., Lamb, M.E., Lewkowicz, K.S., Shoham, R., Dvir, R., & Oxford University Press.
Estes, D. (1985). Security of infant-mother, -father, and Symons, D. (1987). If were all Darwinians, whats the fuss about? In
-metapelet attachments among kibbutz-reared Israeli children. C. Crawford, D. Krebs, & M. Smith (Eds.), Sociobiology and psy-
In I. Bretherton & E. Waters (Eds.), Growing points in attachment chology (pp. 121145). Hillsdale, NJ: Erlbaum.
theory and research (pp. 257575). Oxford, United Kingdom: Thompson, R.A. (2008). Early attachment and later development. In
Blackwell. J. Cassidy & P.R. Shaver, (Eds.), Handbook of attachment: theory,
Schachner, D.A., & Shaver, P.R. (2004). Attachment dimensions and research, and clinical applications (2nd ed., pp. 348365). New
sexual motives. Personal Relationships, 11, 179195. York: Guilford Press.
Schmitt, D.P. (2005). Is short-term mating the maladaptive result of Tidwell, M.O., Reis, H.T., & Shaver, P.R. (1996). Attachment, attrac-
insecure attachment? A test of competing evolutionary perspec- tiveness, and social interaction: A diary study. Journal of Person-
tives. Personality and Social Psychology Bulletin, 31, 747768. ality and Social Psychology, 71, 729745.
Schmitt, D.P., Alcalay, L., Allensworth, M., Allik, J., Ault, L., Tooby, J., & Cosmides, L. (1990). On the universality of human nature
Austers, I., et al. (2004). Patterns and universals of adult romantic and the uniqueness of the individual: The role of genetics and
attachment across 62 cultural regions: Are models of self and of adaptation. Journal of Personality, 58, 1767.
other pancultural constructs? Journal of Cross-Cultural Psychol- Tooke, W., & Camire, L. (1991). Pattern of deception in intrasexual
ogy, 35, 367402. and intersexual mating strategies. Ethology and Sociobiology, 12,
Schwagmeyer, P.L. (1980). Alarm calling behavior of the thirteen- 345364.
lined ground squirrel. Spermophilus tridecemlineatus. Behavioral Townsend, A.L., & Franks, M.M. (1995). Binding ties: Closeness and
Ecology and Sociobiology, 7, 195200. conflict in adult childrens caregiving relationships. Psychology
Shaver, P.R., & Brennan, K.A. (1992). Attachment styles and the Big and Aging, 10, 343351.
Five personality traits: Their connections with each other and van IJzendoorn, M.H. (1995). Adult attachment representations, par-
with romantic relationship outcomes. Personality and Social Psy- ental responsiveness, and infant attachment: A meta-analysis on
chology Bulletin, 18, 536545. the predictive validity of the Adult Attachment Interview. Psycho-
Shaver, P.R., & Clark, C.L. (1994). The psychodynamics of adult logical Bulletin, 117, 387403.
romantic attachment. In J.M. Masling & R.F. Bornstein (Eds.), van IJzendoorn, M.H., Goossens, F.A., Kroonenberg, P.M., &
Empirical perspectives on object relations theories (pp. 105 Tavecchio, L.W.C. (1985). Dependent attachment: B4. Children
156). Washington, DC: American Psychological Association. in the strange situation. Psychological Reports, 57, 439451.
Shaver, P.R., & Mikulincer, M. (2002). Attachment-related psychody- van IJzendoorn, M.H., & Kroonenberg, P.M. (1988). Cross-cultural
namics. Attachment and Human Development, 4, 133161. patterns of attachment: A meta-analysis of the strange situation.
Sherman, P.W. (1977). Nepotism and the evolution of alarm calls. Sci- Child Development, 59, 147156.
ence, 197, 12461253. van IJzendoorn, M.H., & Sagi, A. (1999). Cross-cultural patterns of
Sherman, P.W. (1980). The meaning of nepotism. American Natural- attachment: Universal and cultural dimensions. In J. Cassidy &
ist, 116, 604606. P.R. Shaver (Eds.), Handbook of attachment: Theory, research,
Sherman, P.W. (1981). Kinship, demography, and Beldings ground and clinical applications (pp. 713734). New York: Guilford.
squirrel nepotism. Behavioral Ecology and Sociobiology, 8, Vaughn, B.E., Bost, K.K., & van IJzendoorn, M.H. (2008). Attach-
251259. ment and temperament: Additive and interactive influences on
Sime, J.D. (1983). Affiliative behavior during escape to building exits. behavior, affect, and cognition during infancy and childhood. In
Journal of Environmental Psychology, 3, 2141. J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment: Theory,
Sime, J.D. (1985). Movement toward the familiar: Person and place research, and clinical applications (2nd ed., pp. 192216). New
affiliation in a fire entrapment setting. Environment and Behavior, York: Guilford Press.
17, 697724. Waters, H.S., Rodrigues, L.M., & Ridgeway, D. (1998). Cognitive
Simpson, J.A., & Belsky, J. (2008). Attachment theory in modern evo- underpinnings of narrative attachment assessment. Journal of
lutionary perspective. In J. Cassidy & P.R. Shaver (Eds.), Hand- Experimental Child Psychology, 71, 211234.
book of attachment: Theory, research, and clinical applications Waters, H.S., & Waters, E. (2006). The attachment working models
(2nd ed., pp. 131157). New York: Guilford Press. concept: Among other things, we build script-like representations

140 Downloaded from at Interdisciplinary Center for on May 4, 2010

The Attachment Paradox 141

of secure base experiences. Attachment and Human Development, Wilson, D.S., Van Vugt, M., & OGorman, R. (2008). Multilevel
8, 185197. selection theory and major evolutionary transitions. Current
Wells, H.G. (1898). The war of the worlds. London: Heinemann. Directions in Psychological Science, 17, 69.
West-Eberhard, M.J. (2003). Developmental plasticity and evolution. Wilson, D.S., & Wilson, E.D. (2007). Rethinking the theoretical
New York: Oxford University Press. founding of sociobiology. Quarterly Review of Biology, 82,
Williams, N.L., & Riskind, J.H. (2004). Adult romantic attachment 327348.
and cognitive vulnerabilities to anxiety and depression: Examining Zevalkink, J., Riksen-Walraven, J.M.A., & Van-Lieshout, C.F.M.
the interpersonal basis of vulnerability models. Journal of (1999). Attachment in the Indonesian caregiving context. Social
Cognitive Psychotherapy: An International Quarterly, 18, 224. Development, 8, 2140.

Downloaded from at Interdisciplinary Center for on May 4, 2010 141

View publication stats